Google Cloud Scheduler 在 Function 成功时调用 Google Cloud Function（返回 200 状态）仍然获得 500 状态

Question

我有一个调度程序作业每小时调用一次我的函数。该函数工作正常，每次都返回 HTTP 200 状态。但是，每 3-5 次调用，调度程序作业 returns 一个 HTTP 500 状态。这个问题的cause/fix是什么？

示例：函数日志显示的内容如下：

D 2020-02-03T10:02:10.958520185Z Function execution took 130785 ms, finished with status code: 200 
D 2020-02-03T11:01:40.819608573Z Function execution took 99762 ms, finished with status code: 200 
D 2020-02-03T12:01:41.049430737Z Function execution took 100126 ms, finished with status code: 200 
D 2020-02-03T13:02:07.369401657Z Function execution took 127213 ms, finished with status code: 200 
D 2020-02-03T14:04:24.352839424Z Function execution took 263896 ms, finished with status code: 200 
D 2020-02-03T15:03:14.664760657Z Function execution took 194125 ms, finished with status code: 200 
D 2020-02-03T16:06:23.162542969Z Function execution took 382609 ms, finished with status code: 200 
D 2020-02-03T17:03:17.458640891Z Function execution took 196799 ms, finished with status code: 200 
D 2020-02-03T18:02:54.614556691Z Function execution took 170119 ms, finished with status code: 200 
D 2020-02-03T19:04:43.064083790Z Function execution took 277775 ms, finished with status code: 200 
D 2020-02-03T20:02:59.315497864Z Function execution took 178499 ms, finished with status code: 200

这些是调度程序日志中的示例

2020-02-03 11:03:00.567 CST
{"jobName":"projects/my-project/locations/us-west2/jobs/my-function-trigger","targetType":"HTTP","url":"https://us-central1-my-function.cloudfunctions.net/analytics-archiver","status":"UNKNOWN","@type":"type.googleapis.com/google.cloud.scheduler.logging.AttemptFinished"}
{
 httpRequest: {
 }
 insertId: "redacted"  
 jsonPayload: {
  @type: "type.googleapis.com/google.cloud.scheduler.logging.AttemptFinished"   
  jobName: "projects/my-project/locations/us-west2/jobs/my-function-trigger"   
  status: "UNKNOWN"   
  targetType: "HTTP"   
  url: "https://us-central1-my-project.cloudfunctions.net/my-function"   
 }
 logName: "projects/my-project/logs/cloudscheduler.googleapis.com%2Fexecutions"  
 receiveTimestamp: "2020-02-03T17:03:00.567786781Z"  
 resource: {
  labels: {
   job_id: "my-function-trigger"    
   location: "us-west2"    
   project_id: "my-function"    
  }
  type: "cloud_scheduler_job"   
 }
 severity: "ERROR"  
 timestamp: "2020-02-03T17:03:00.567786781Z"  
}


2020-02-03 13:03:00.765 CST
{"jobName":"projects/my-project/locations/us-west2/jobs/my-function-trigger","targetType":"HTTP","url":"https://us-central1-my-project.cloudfunctions.net/my-function","status":"UNKNOWN","@type":"type.googleapis.com/google.cloud.scheduler.logging.AttemptFinished"}
Expand all | Collapse all{
 httpRequest: {
 }
 insertId: "redacted"  
 jsonPayload: {
  @type: "type.googleapis.com/google.cloud.scheduler.logging.AttemptFinished"   
  jobName: "projects/my-project/locations/us-west2/jobs/my-function-trigger"   
  status: "UNKNOWN"   
  targetType: "HTTP"   
  url: "https://us-central1-my-project.cloudfunctions.net/my-function"   
 }
 logName: "projects/my-project/logs/cloudscheduler.googleapis.com%2Fexecutions"  
 receiveTimestamp: "2020-02-03T19:03:00.765993324Z"  
 resource: {
  labels: {
   job_id: "my-function-trigger"    
   location: "us-west2"    
   project_id: "my-project"    
  }
  type: "cloud_scheduler_job"   
 }
 severity: "ERROR"  
 timestamp: "2020-02-03T19:03:00.765993324Z"  
}

Answer 1

考虑到这些请求所花费的时间，这似乎导致 Google 调度程序超时，最终导致某些调用失败。

根据文档 cron.yaml Reference，您可以在那里配置调度程序在超时前花费的最长时间。我建议您看一下它并确认您的配置方式，并尝试将您的调用保持在 cron.yaml 文件中设置的时间内。

如果信息对您有所帮助，请告诉我！

Answer 2

我遇到了完全相同的问题。虽然您在控制台中还看不到它，但 Google Cloud Scheduler 有一些标志，您可以通过 gcloud scheduler jobs create http 命令设置。这是我使用的示例：

gcloud scheduler jobs create http my-job \
  --schedule="0 * * * *" \
  --uri=https://europe-west1-${PROJECT_ID}.cloudfunctions.net/my-func/ \
  --http-method=POST \
  --message-body="my-body" \
  --max-retry-attempts 3 \
  --max-backoff 10s \
  --attempt-deadline 10m

特别是当运行功能持续几分钟而不是几秒钟时，尝试截止日期似乎很重要。设置这些标志减轻了我的一些问题，但不是全部。有关更多标志，请参阅 documentation。

此外，如果这没有帮助，则可能是 Google 背后某处的服务器端错误。这是由于 UNKNOWN 错误状态，您可以在此 table 中查找。我自己得到了内部错误状态，也称为服务器端错误。 Google..

不是特别有用

Google Cloud Scheduler 在 Function 成功时调用 Google Cloud Function（返回 200 状态）仍然获得 500 状态

Google Cloud Scheduler calling Google Cloud Function when Function has succeeded (returned 200 status) still gets 500 status

google-cloud-platform

google-cloud-functions

google-cloud-scheduler